Training Excitatory-Inhibitory Recurrent Neural Networks for Cognitive Tasks: A Simple and Flexible Framework

Brain-like border ownership signals support prediction of natural videos

To make sense of visual scenes, the brain must segment foreground from background. This is thought to be facilitated by neurons that signal border ownership (BOS), which indicate which side of a border in their receptive field is owned by an object. How these signals emerge without a teaching signal of what is foreground remains unclear. Here we find that many units in PredNet, a self-supervised deep neural network trained to predict future frames in natural videos, are selective for BOS. They share key properties with BOS neurons in the brain, including robustness to object transformations and hysteresis. Ablation revealed that BOS units contribute more to prediction than other units for videos with moving objects. Our findings suggest that BOS neurons might emerge due to an evolutionary or developmental pressure to predict future input in natural, complex dynamic environments, even without an explicit requirement to segment foreground from background.

Phase of firing does not reflect temporal order in sequence memory of humans and recurrent neural networks

The temporal order of a sequence of events has been thought to be reflected in the ordered firing of neurons at different phases of theta oscillations. Here we assess this by measuring single neuron activity (1,420 neurons) and local field potentials (921 channels) in the medial temporal lobe of 16 patients with epilepsy performing a working-memory task for temporal order. During memory maintenance, we observe theta oscillations, preferential firing of single neurons to theta phase and a close relationship between phase of firing and item position. However, the firing order did not match item order. Training recurrent neural networks to perform an analogous task, we also show the generation of theta oscillations, theta phase-dependent firing related to item position and, again, no match between firing and item order. Rather, our results suggest a mechanistic link between phase order, stimulus timing and oscillation frequency. In both biological and artificial neural networks, we provide evidence supporting the role of phase of firing in working-memory processing.

Mediodorsal thalamus regulates task uncertainty to enable cognitive flexibility

The mediodorsal (MD) thalamus is a critical partner for the prefrontal cortex (PFC) in cognitive control. Accumulating evidence has shown that the MD regulates task uncertainty in decision making and enhance cognitive flexibility. However, the computational mechanism of this cognitive process remains unclear. Here we trained biologically-constrained computational models to delineate the mechanistic role of MD in context-dependent decision making. We show that the addition of a feedforward MD structure to the recurrent PFC increases robustness to low cueing signal-to-noise ratio, enhances working memory, and enables rapid context switching. Incorporating genetically identified thalamocortical connectivity and interneuron cell types into the model replicates key neurophysiological findings in task-performing animals. Our model reveals computational mechanisms and geometric interpretations of MD in regulating cue uncertainty and context switching to enable cognitive flexibility. Our model makes experimentally testable predictions linking cognitive deficits with disrupted thalamocortical connectivity, prefrontal excitation-inhibition imbalance and dysfunctional inhibitory cell types.

A Neural Circuit Framework for Economic Choice: From Building Blocks of Valuation to Compositionality in Multitasking

Value-guided decisions are at the core of reinforcement learning and neuroeconomics, yet the basic computations they require remain poorly understood at the mechanistic level. For instance, how does the brain implement the multiplication of reward magnitude by probability to yield an expected value? Where within a neural circuit is the indifference point for comparing reward types encoded? How do learned values generalize to novel options? Here, we introduce a biologically plausible model that adheres to Dale's law and is trained on five choice tasks, offering potential answers to these questions. The model captures key neurophysiological observations from the orbitofrontal cortex of monkeys and generalizes to novel offer values. Using a single network model to solve diverse tasks, we identified compositional neural representations-quantified via task variance analysis and corroborated by curriculum learning. This work provides testable predictions that probe the neural basis of decision making and its disruption in neuropsychiatric disorders.

Latent circuit inference from heterogeneous neural responses during cognitive tasks

Higher cortical areas carry a wide range of sensory, cognitive and motor signals mixed in heterogeneous responses of single neurons tuned to multiple task variables. Dimensionality reduction methods that rely on correlations between neural activity and task variables leave unknown how heterogeneous responses arise from connectivity to drive behavior. We develop the latent circuit model, a dimensionality reduction approach in which task variables interact via low-dimensional recurrent connectivity to produce behavioral output. We apply the latent circuit inference to recurrent neural networks trained to perform a context-dependent decision-making task and find a suppression mechanism in which contextual representations inhibit irrelevant sensory responses. We validate this mechanism by confirming the behavioral effects of patterned connectivity perturbations predicted by the latent circuit model. We find similar suppression of irrelevant sensory responses in the prefrontal cortex of monkeys performing the same task. We show that incorporating causal interactions among task variables is critical for identifying behaviorally relevant computations from neural response data.

Humans actively reconfigure neural task states

The ability to switch between tasks is a core component of adaptive cognition, yet a mechanistic understanding of this capacity has remained elusive. In particular, there are longstanding debates over the extent to which task-switching is primarily influenced by interference from previous tasks or active preparation for upcoming tasks. We advance this debate by modeling the evolution of task representations in human electroencephalographic recordings as linear dynamical systems over a high-dimensional latent space. Using control theoretic analyses of dynamical systems fit to both brains and task-optimized neural networks, we reveal neural signatures of active preparation that reconcile the role of interference and control in task switching. These findings inform a core debate in cognitive control and offer a promising new paradigm for human neuroimaging analysis.

Recurrent neural networks with transient trajectory explain working memory encoding mechanisms

Whether working memory (WM) is encoded by persistent activity using attractors or by dynamic activity using transient trajectories has been debated for decades in both experimental and modeling studies, and a consensus has not been reached. Even though many recurrent neural networks (RNNs) have been proposed to simulate WM, most networks are designed to match respective experimental observations and show either transient or persistent activities. Those few which consider networks with both activity patterns have not attempted to directly compare their memory capabilities. In this study, we build transient-trajectory-based RNNs (TRNNs) and compare them to vanilla RNNs with more persistent activities. The TRNN incorporates biologically plausible modifications, including self-inhibition, sparse connection and hierarchical topology. Besides activity patterns resembling animal recordings and retained versatility to variable encoding time, TRNNs show better performance in delayed choice and spatial memory reinforcement learning tasks. Therefore, this study provides evidence supporting the transient activity theory to explain the WM mechanism from the model designing point of view.

Effects of noise and metabolic cost on cortical task representations

Cognitive flexibility requires both the encoding of task-relevant and the ignoring of task-irrelevant stimuli. While the neural coding of task-relevant stimuli is increasingly well understood, the mechanisms for ignoring task-irrelevant stimuli remain poorly understood. Here, we study how task performance and biological constraints jointly determine the coding of relevant and irrelevant stimuli in neural circuits. Using mathematical analyses and task-optimized recurrent neural networks, we show that neural circuits can exhibit a range of representational geometries depending on the strength of neural noise and metabolic cost. By comparing these results with recordings from primate prefrontal cortex (PFC) over the course of learning, we show that neural activity in PFC changes in line with a minimal representational strategy. Specifically, our analyses reveal that the suppression of dynamically irrelevant stimuli is achieved by activity-silent, sub-threshold dynamics. Our results provide a normative explanation as to why PFC implements an adaptive, minimal representational strategy.

Random noise promotes slow heterogeneous synaptic dynamics important for robust working memory computation

Recurrent neural networks (RNNs) based on model neurons that communicate via continuous signals have been widely used to study how cortical neural circuits perform cognitive tasks. Training such networks to perform tasks that require information maintenance over a brief period (i.e., working memory tasks) remains a challenge. Inspired by the robust information maintenance observed in higher cortical areas such as the prefrontal cortex, despite substantial inherent noise, we investigated the effects of random noise on RNNs across different cognitive functions, including working memory. Our findings reveal that random noise not only speeds up training but also enhances the stability and performance of RNNs on working memory tasks. Importantly, this robust working memory performance induced by random noise during training is attributed to an increase in synaptic decay time constants of inhibitory units, resulting in slower decay of stimulus-specific activity critical for memory maintenance. Our study reveals the critical role of noise in shaping neural dynamics and cognitive functions, suggesting that inherent variability may be a fundamental feature driving the specialization of inhibitory neurons to support stable information processing in higher cortical regions.

Unconditional stability of a recurrent neural circuit implementing divisive normalization

Stability in recurrent neural models poses a significant challenge, particularly in developing biologically plausible neurodynamical models that can be seamlessly trained. Traditional cortical circuit models are notoriously difficult to train due to expansive nonlinearities in the dynamical system, leading to an optimization problem with nonlinear stability constraints that are difficult to impose. Conversely, recurrent neural networks (RNNs) excel in tasks involving sequential data but lack biological plausibility and interpretability. In this work, we address these challenges by linking dynamic divisive normalization (DN) to the stability of "oscillatory recurrent gated neural integrator circuits" (ORGaNICs), a biologically plausible recurrent cortical circuit model that dynamically achieves DN and that has been shown to simulate a wide range of neurophysiological phenomena. By using the indirect method of Lyapunov, we prove the remarkable property of unconditional local stability for an arbitrary-dimensional ORGaNICs circuit when the recurrent weight matrix is the identity. We thus connect ORGaNICs to a system of coupled damped harmonic oscillators, which enables us to derive the circuit's energy function, providing a normative principle of what the circuit, and individual neurons, aim to accomplish. Further, for a generic recurrent weight matrix, we prove the stability of the 2D model and demonstrate empirically that stability holds in higher dimensions. Finally, we show that ORGaNICs can be trained by backpropagation through time without gradient clipping/scaling, thanks to its intrinsic stability property and adaptive time constants, which address the problems of exploding, vanishing, and oscillating gradients. By evaluating the model's performance on RNN benchmarks, we find that ORGaNICs outperform alternative neurodynamical models on static image classification tasks and perform comparably to LSTMs on sequential tasks.

CONSTRUCTING BIOLOGICALLY CONSTRAINED RNNS VIA DALE'S BACKPROP AND TOPOLOGICALLY-INFORMED PRUNING

Recurrent neural networks (RNNs) have emerged as a prominent tool for modeling cortical function, and yet their conventional architecture is lacking in physiological and anatomical fidelity. In particular, these models often fail to incorporate two crucial biological constraints: i) Dale's law, i.e., sign constraints that preserve the "type" of projections from individual neurons, and ii) Structured connectivity motifs, i.e., highly sparse yet defined connections amongst various neuronal populations. Both constraints are known to impair learning performance in artificial neural networks, especially when trained to perform complicated tasks; but as modern experimental methodologies allow us to record from diverse neuronal populations spanning multiple brain regions, using RNN models to study neuronal interactions without incorporating these fundamental biological properties raises questions regarding the validity of the insights gleaned from them. To address these concerns, our work develops methods that let us train RNNs which respect Dale's law whilst simultaneously maintaining a specific sparse connectivity pattern across the entire network. We provide mathematical grounding and guarantees for our approaches incorporating both types of constraints, and show empirically that our models match the performance of RNNs trained without any constraints. Finally, we demonstrate the utility of our methods for inferring multi-regional interactions by training RNN models of the cortical network to reconstruct 2-photon calcium imaging data during visual behaviour in mice, whilst enforcing data-driven, cell-type specific connectivity constraints between various neuronal populations spread across multiple cortical layers and brain areas. In doing so, we find that the interactions inferred by our model corroborate experimental findings in agreement with the theory of predictive coding, thus validating the applicability of our methods.

Dynamic tuning of neural stability for cognitive control

The brain is thought to execute cognitive control by actively maintaining and flexibly updating patterns of neural activity that represent goals and rules. However, while actively maintaining patterns of activity requires robustness against noise and distractors, updating the activity requires sensitivity to task-relevant inputs. How these conflicting demands can be reconciled in a single neural system remains unclear. Here, we study the prefrontal cortex of monkeys maintaining a covert rule and integrating sensory inputs toward a choice. Following the onset of neural responses, sensory integration evolves with a 70 ms delay. Using a stability analysis and a recurrent neural network model trained to perform the task, we show that this delay enables a transient, system-level destabilization, opening a temporal window to selectively incorporate new information. This mechanism allows robustness and sensitivity to coexist in a neural system and hierarchically updates patterns of neural activity, providing a general framework for cognitive control. Furthermore, it reveals a learned, explicit rule representation, suggesting a reconciliation between the symbolic and connectionist approaches for building intelligent machines.

Brain-inspired multisensory integration neural network for cross-modal recognition through spatiotemporal dynamics and deep learning

The integration and interaction of cross-modal senses in brain neural networks can facilitate high-level cognitive functionalities. In this work, we proposed a bioinspired multisensory integration neural network (MINN) that integrates visual and audio senses for recognizing multimodal information across different sensory modalities. This deep learning-based model incorporates a cascading framework of parallel convolutional neural networks (CNNs) for extracting intrinsic features from visual and audio inputs, and a recurrent neural network (RNN) for multimodal information integration and interaction. The network was trained using synthetic training data generated for digital recognition tasks. It was revealed that the spatial and temporal features extracted from visual and audio inputs by CNNs were encoded in subspaces orthogonal with each other. In integration epoch, network state evolved along quasi-rotation-symmetric trajectories and a structural manifold with stable attractors was formed in RNN, supporting accurate cross-modal recognition. We further evaluated the robustness of the MINN algorithm with noisy inputs and asynchronous digital inputs. Experimental results demonstrated the superior performance of MINN for flexible integration and accurate recognition of multisensory information with distinct sense properties. The present results provide insights into the computational principles governing multisensory integration and a comprehensive neural network model for brain-inspired intelligence.

Dynamical mechanisms of how an RNN keeps a beat, uncovered with a low-dimensional reduced model

Despite music's omnipresence, the specific neural mechanisms responsible for perceiving and anticipating temporal patterns in music are unknown. To study potential mechanisms for keeping time in rhythmic contexts, we train a biologically constrained RNN, with excitatory (E) and inhibitory (I) units, on seven different stimulus tempos (2-8 Hz) on a synchronization and continuation task, a standard experimental paradigm. Our trained RNN generates a network oscillator that uses an input current (context parameter) to control oscillation frequency and replicates key features of neural dynamics observed in neural recordings of monkeys performing the same task. We develop a reduced three-variable rate model of the RNN and analyze its dynamic properties. By treating our understanding of the mathematical structure for oscillations in the reduced model as predictive, we confirm that the dynamical mechanisms are found also in the RNN. Our neurally plausible reduced model reveals an E-I circuit with two distinct inhibitory sub-populations, of which one is tightly synchronized with the excitatory units.

Adaptive robustness through incoherent signaling mechanisms in a regenerative brain

Animal behavior emerges from collective dynamics of neurons, making it vulnerable to damage. Paradoxically, many organisms exhibit a remarkable ability to maintain significant behavior even after large-scale neural injury. Molecular underpinnings of this extreme robustness remain largely unknown. Here, we develop a quantitative pipeline to measure long-lasting latent states in planarian flatworm behaviors during whole-brain regeneration. By combining >20,000 animal trials with neural network modeling, we show that long-range volumetric peptidergic signals allow the planarian to rapidly restore coarse behavior output after large perturbations to the nervous system, while slow restoration of small-molecule neuromodulator functions refines precision. This relies on the different time and length scales of neuropeptide and small-molecule transmission to generate incoherent patterns of neural activity that competitively regulate behavior. Controlling behavior through opposing communication mechanisms creates a more robust system than either alone and may serve as a generalizable approach for constructing robust neural networks.

Flexible gating between subspaces in a neural network model of internally guided task switching

Behavioral flexibility relies on the brain's ability to switch rapidly between multiple tasks, even when the task rule is not explicitly cued but must be inferred through trial and error. The underlying neural circuit mechanism remains poorly understood. We investigated recurrent neural networks (RNNs) trained to perform an analog of the classic Wisconsin Card Sorting Test. The networks consist of two modules responsible for rule representation and sensorimotor mapping, respectively, where each module is comprised of a circuit with excitatory neurons and three major types of inhibitory neurons. We found that rule representation by self-sustained persistent activity across trials, error monitoring and gated sensorimotor mapping emerged from training. Systematic dissection of trained RNNs revealed a detailed circuit mechanism that is consistent across networks trained with different hyperparameters. The networks' dynamical trajectories for different rules resided in separate subspaces of population activity; the subspaces collapsed and performance was reduced to chance level when dendrite-targeting somatostatin-expressing interneurons were silenced, illustrating how a phenomenological description of representational subspaces is explained by a specific circuit mechanism.

Spiking mode-based neural networks

Spiking neural networks play an important role in brainlike neuromorphic computations and in studying working mechanisms of neural circuits. One drawback of training a large-scale spiking neural network is that updating all weights is quite expensive. Furthermore, after training, all information related to the computational task is hidden into the weight matrix, prohibiting us from a transparent understanding of circuit mechanisms. Therefore, in this work, we address these challenges by proposing a spiking mode-based training protocol, where the recurrent weight matrix is explained as a Hopfield-like multiplication of three matrices: input modes, output modes, and a score matrix. The first advantage is that the weight is interpreted by input and output modes and their associated scores characterizing the importance of each decomposition term. The number of modes is thus adjustable, allowing more degrees of freedom for modeling the experimental data. This significantly reduces the training cost because of significantly reduced space complexity for learning. Training spiking networks is thus carried out in the mode-score space. The second advantage is that one can project the high-dimensional neural activity (filtered spike train) in the state space onto the mode space which is typically of a low dimension, e.g., a few modes are sufficient to capture the shape of the underlying neural manifolds. We successfully apply our framework in two computational tasks-digit classification and selective sensory integration tasks. Our method thus accelerates the training of spiking neural networks by a Hopfield-like decomposition, and moreover this training leads to low-dimensional attractor structures of high-dimensional neural dynamics.

Neural representational geometries reflect behavioral differences in monkeys and recurrent neural networks

Animals likely use a variety of strategies to solve laboratory tasks. Traditionally, combined analysis of behavioral and neural recording data across subjects employing different strategies may obscure important signals and give confusing results. Hence, it is essential to develop techniques that can infer strategy at the single-subject level. We analyzed an experiment in which two male monkeys performed a visually cued rule-based task. The analysis of their performance shows no indication that they used a different strategy. However, when we examined the geometry of stimulus representations in the state space of the neural activities recorded in dorsolateral prefrontal cortex, we found striking differences between the two monkeys. Our purely neural results induced us to reanalyze the behavior. The new analysis showed that the differences in representational geometry are associated with differences in the reaction times, revealing behavioral differences we were unaware of. All these analyses suggest that the monkeys are using different strategies. Finally, using recurrent neural network models trained to perform the same task, we show that these strategies correlate with the amount of training, suggesting a possible explanation for the observed neural and behavioral differences.

On the neural networks of self and other bias and their role in emergent social interactions

Extensive research has documented the brain networks that play an integral role in bias, or the alteration and filtration of information processing in a manner that fundamentally favors an individual. The roots of bias, whether self- or other-oriented, are a complex constellation of neural and psychological processes that start at the most fundamental levels of sensory processing. From the millisecond information is received in the brain it is filtered at various levels and through various brain networks in relation to extant intrinsic activity to provide individuals with a perception of reality that complements and satisfies the conscious perceptions they have for themselves and the cultures in which they were reared. The products of these interactions, in turn, are dynamically altered by the introduction of others, be they friends or strangers who are similar or different in socially meaningful ways. While much is known about the various ways that basic biases alter specific aspects of neural function to support various forms of bias, the breadth and scope of the phenomenon remains entirely unclear. The purpose of this review is to examine the brain networks that shape (i.e., bias) the self-concept and how interactions with similar (ingroup) compared to dissimilar (outgroup) others alter these network (and subsequent interpersonal) interactions in fundamental ways. Throughout, focus is placed on an emerging understanding of the brain as a complex system, which suggests that many of these network interactions likely occur on a non-linear scale that blurs the lines between network hierarchies.

A multi-region recurrent circuit for evidence accumulation in rats

Decision-making based on noisy evidence requires accumulating evidence and categorizing it to form a choice. Here we evaluate a proposed feedforward and modular mapping of this process in rats: evidence accumulated in anterodorsal striatum (ADS) is categorized in prefrontal cortex (frontal orienting fields, FOF). Contrary to this, we show that both regions appear to be indistinguishable in their encoding/decoding of accumulator value and communicate this information bidirectionally. Consistent with a role for FOF in accumulation, silencing FOF to ADS projections impacted behavior throughout the accumulation period, even while nonselective FOF silencing did not. We synthesize these findings into a multi-region recurrent neural network trained with a novel approach. In-silico experiments reveal that multiple scales of recurrence in the cortico-striatal circuit rescue computation upon nonselective FOF perturbations. These results suggest that ADS and FOF accumulate evidence in a recurrent and distributed manner, yielding redundant representations and robustness to certain perturbations.

The computational foundations of dynamic coding in working memory

Working memory (WM) is a fundamental aspect of cognition. WM maintenance is classically thought to rely on stable patterns of neural activities. However, recent evidence shows that neural population activities during WM maintenance undergo dynamic variations before settling into a stable pattern. Although this has been difficult to explain theoretically, neural network models optimized for WM typically also exhibit such dynamics. Here, we examine stable versus dynamic coding in neural data, classical models, and task-optimized networks. We review principled mathematical reasons for why classical models do not, while task-optimized models naturally do exhibit dynamic coding. We suggest an update to our understanding of WM maintenance, in which dynamic coding is a fundamental computational feature rather than an epiphenomenon.

Flexible gating between subspaces in a neural network model of internally guided task switching

Behavioral flexibility relies on the brain's ability to switch rapidly between multiple tasks, even when the task rule is not explicitly cued but must be inferred through trial and error. The underlying neural circuit mechanism remains poorly understood. We investigated recurrent neural networks (RNNs) trained to perform an analog of the classic Wisconsin Card Sorting Test. The networks consist of two modules responsible for rule representation and sensorimotor mapping, respectively, where each module is comprised of a circuit with excitatory neurons and three major types of inhibitory neurons. We found that rule representation by self-sustained persistent activity across trials, error monitoring and gated sensorimotor mapping emerged from training. Systematic dissection of trained RNNs revealed a detailed circuit mechanism that is consistent across networks trained with different hyperparameters. The networks' dynamical trajectories for different rules resided in separate subspaces of population activity; the subspaces collapsed and performance was reduced to chance level when dendrite-targeting somatostatin-expressing interneurons were silenced, illustrating how a phenomenological description of representational subspaces is explained by a specific circuit mechanism.

Effect in the spectra of eigenvalues and dynamics of RNNs trained with excitatory-inhibitory constraint

In order to comprehend and enhance models that describes various brain regions it is important to study the dynamics of trained recurrent neural networks. Including Dale's law in such models usually presents several challenges. However, this is an important aspect that allows computational models to better capture the characteristics of the brain. Here we present a framework to train networks using such constraint. Then we have used it to train them in simple decision making tasks. We characterized the eigenvalue distributions of the recurrent weight matrices of such networks. Interestingly, we discovered that the non-dominant eigenvalues of the recurrent weight matrix are distributed in a circle with a radius less than 1 for those whose initial condition before training was random normal and in a ring for those whose initial condition was random orthogonal. In both cases, the radius does not depend on the fraction of excitatory and inhibitory units nor the size of the network. Diminution of the radius, compared to networks trained without the constraint, has implications on the activity and dynamics that we discussed here.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11571-023-09956-w.

Approximating Nonlinear Functions With Latent Boundaries in Low-Rank Excitatory-Inhibitory Spiking Networks

Deep feedforward and recurrent neural networks have become successful functional models of the brain, but they neglect obvious biological details such as spikes and Dale's law. Here we argue that these details are crucial in order to understand how real neural circuits operate. Towards this aim, we put forth a new framework for spike-based computation in low-rank excitatory-inhibitory spiking networks. By considering populations with rank-1 connectivity, we cast each neuron's spiking threshold as a boundary in a low-dimensional input-output space. We then show how the combined thresholds of a population of inhibitory neurons form a stable boundary in this space, and those of a population of excitatory neurons form an unstable boundary. Combining the two boundaries results in a rank-2 excitatory-inhibitory (EI) network with inhibition-stabilized dynamics at the intersection of the two boundaries. The computation of the resulting networks can be understood as the difference of two convex functions and is thereby capable of approximating arbitrary non-linear input-output mappings. We demonstrate several properties of these networks, including noise suppression and amplification, irregular activity and synaptic balance, as well as how they relate to rate network dynamics in the limit that the boundary becomes soft. Finally, while our work focuses on small networks (5-50 neurons), we discuss potential avenues for scaling up to much larger networks. Overall, our work proposes a new perspective on spiking networks that may serve as a starting point for a mechanistic understanding of biological spike-based computation.

Exploring Flip Flop memories and beyond: training Recurrent Neural Networks with key insights

Training neural networks to perform different tasks is relevant across various disciplines. In particular, Recurrent Neural Networks (RNNs) are of great interest in Computational Neuroscience. Open-source frameworks dedicated to Machine Learning, such as Tensorflow and Keras have produced significant changes in the development of technologies that we currently use. This work contributes by comprehensively investigating and describing the application of RNNs for temporal processing through a study of a 3-bit Flip Flop memory implementation. We delve into the entire modeling process, encompassing equations, task parametrization, and software development. The obtained networks are meticulously analyzed to elucidate dynamics, aided by an array of visualization and analysis tools. Moreover, the provided code is versatile enough to facilitate the modeling of diverse tasks and systems. Furthermore, we present how memory states can be efficiently stored in the vertices of a cube in the dimensionally reduced space, supplementing previous results with a distinct approach.

Synaptic wiring motifs in posterior parietal cortex support decision-making

The posterior parietal cortex exhibits choice-selective activity during perceptual decision-making tasks1-10. However, it is not known how this selective activity arises from the underlying synaptic connectivity. Here we combined virtual-reality behaviour, two-photon calcium imaging, high-throughput electron microscopy and circuit modelling to analyse how synaptic connectivity between neurons in the posterior parietal cortex relates to their selective activity. We found that excitatory pyramidal neurons preferentially target inhibitory interneurons with the same selectivity. In turn, inhibitory interneurons preferentially target pyramidal neurons with opposite selectivity, forming an opponent inhibition motif. This motif was present even between neurons with activity peaks in different task epochs. We developed neural-circuit models of the computations performed by these motifs, and found that opponent inhibition between neural populations with opposite selectivity amplifies selective inputs, thereby improving the encoding of trial-type information. The models also predict that opponent inhibition between neurons with activity peaks in different task epochs contributes to creating choice-specific sequential activity. These results provide evidence for how synaptic connectivity in cortical circuits supports a learned decision-making task.

How connectivity structure shapes rich and lazy learning in neural circuits

In theoretical neuroscience, recent work leverages deep learning tools to explore how some network attributes critically influence its learning dynamics. Notably, initial weight distributions with small (resp. large) variance may yield a rich (resp. lazy) regime, where significant (resp. minor) changes to network states and representation are observed over the course of learning. However, in biology, neural circuit connectivity could exhibit a low-rank structure and therefore differs markedly from the random initializations generally used for these studies. As such, here we investigate how the structure of the initial weights -- in particular their effective rank -- influences the network learning regime. Through both empirical and theoretical analyses, we discover that high-rank initializations typically yield smaller network changes indicative of lazier learning, a finding we also confirm with experimentally-driven initial connectivity in recurrent neural networks. Conversely, low-rank initialization biases learning towards richer learning. Importantly, however, as an exception to this rule, we find lazier learning can still occur with a low-rank initialization that aligns with task and data statistics. Our research highlights the pivotal role of initial weight structures in shaping learning regimes, with implications for metabolic costs of plasticity and risks of catastrophic forgetting.

Connectome-based reservoir computing with the conn2res toolbox

The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses high-dimensional, nonlinear dynamical systems to perform computations and approximate cognitive functions. Here we present conn2res: an open-source Python toolbox for implementing biological neural networks as artificial neural networks. conn2res is modular, allowing arbitrary network architecture and dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from spiking neurons to memristive dynamics. The versatility of the conn2res toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, conn2res sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.

Modelling and Controlling System Dynamics of the Brain: An Intersection of Machine Learning and Control Theory

The human brain, as a complex system, has long captivated multidisciplinary researchers aiming to decode its intricate structure and function. This intricate network has driven scientific pursuits to advance our understanding of cognition, behavior, and neurological disorders by delving into the complex mechanisms underlying brain function and dysfunction. Modelling brain dynamics using machine learning techniques deepens our comprehension of brain dynamics from a computational perspective. These computational models allow researchers to simulate and analyze neural interactions, facilitating the identification of dysfunctions in connectivity or activity patterns. Additionally, the trained dynamical system, serving as a surrogate model, optimizes neurostimulation strategies under the guidelines of control theory. In this chapter, we discuss the recent studies on modelling and controlling brain dynamics at the intersection of machine learning and control theory, providing a framework to understand and improve cognitive function, and treat neurological and psychiatric disorders.

Low-frequency electroacupuncture exerts antinociceptive effects through activation of POMC neural circuit induced endorphinergic input to the periaqueductal gray from the arcuate nucleus

It has been widely recognized that electroacupuncture (EA) inducing the release of β-endorphin represents a crucial mechanism of EA analgesia. The arcuate nucleus (ARC) in the hypothalamus is a vital component of the endogenous opioid peptide system. Serving as an integration center, the periaqueductal gray (PAG) receives neural fiber projections from the frontal cortex, insular cortex, and ARC. However, the specific mechanisms how EA facilitates the release of β-endorphin within the ARC, eliciting analgesic effects are yet to be elucidated. In this study, we conducted in vivo and in vitro experiments by transcriptomics, microdialysis, photogenetics, chemical genetics, and calcium imaging, combined with transgenic animals. Firstly, we detected 2 Hz EA at the Zusanli (ST36) increased the level of β-endorphin and transcriptional level of proopiomelanocortin (POMC). Our transcriptomics profiling demonstrated that 2 Hz EA at the ST36 modulates the expression of c-Fos and Jun B in ARC brain nuclear cluster, and the transcriptional regulation of 2 Hz EA mainly occur in POMC neurons by Immunofluorescence staining verification. Meaning while, 2 Hz EA specifically activated the cAMP-PKA-CREB signaling pathway in ARC which mediating the c-Fos and Jun B transcription, and 2 Hz EA analgesia is dependent on the activation of cAMP-PKA-CREB signaling pathway in ARC. In order to investigate how the β-endorphin produced in ARC transfer to integration center PAG, transneuronal tracing technology was used to observe the 2 Hz EA promoted the neural projection from ARC to PAG compared to 100 Hz EA and sham mice. Inhibited PAGGABA neurons, the transfer of β-endorphin from the ARC nucleus to the PAG nucleus through the ARCPOMC-PAGGABA neural circuit. Furthermore, by manipulating the excitability of POMC neurons from ARCPOMC to PAGGABA using inhibitory chemogenetics and optogenetics, we found that this inhibition significantly reduced transfer of β-endorphin from the ARC nucleus to the PAG nucleus and the effectiveness of 2 Hz EA analgesia in neurological POMC cyclization recombination enzyme (Cre) mice and C57BL/6J mice, which indicates that the transfer of β-endorphin depends on the activation of POMC neurons prefect from ARCPOMC to PAGGABA. These findings contribute to our understanding of the neural circuitry underlying the EA pain-relieving effects and maybe provide valuable insights for optimizing EA stimulation parameters in clinical pain treatment using the in vivo dynamic visual investigating the central analgesic mechanism.

BrainPy, a flexible, integrative, efficient, and extensible framework for general-purpose brain dynamics programming

Elucidating the intricate neural mechanisms underlying brain functions requires integrative brain dynamics modeling. To facilitate this process, it is crucial to develop a general-purpose programming framework that allows users to freely define neural models across multiple scales, efficiently simulate, train, and analyze model dynamics, and conveniently incorporate new modeling approaches. In response to this need, we present BrainPy. BrainPy leverages the advanced just-in-time (JIT) compilation capabilities of JAX and XLA to provide a powerful infrastructure tailored for brain dynamics programming. It offers an integrated platform for building, simulating, training, and analyzing brain dynamics models. Models defined in BrainPy can be JIT compiled into binary instructions for various devices, including Central Processing Unit, Graphics Processing Unit, and Tensor Processing Unit, which ensures high-running performance comparable to native C or CUDA. Additionally, BrainPy features an extensible architecture that allows for easy expansion of new infrastructure, utilities, and machine-learning approaches. This flexibility enables researchers to incorporate cutting-edge techniques and adapt the framework to their specific needs.

Optimal information loading into working memory explains dynamic coding in the prefrontal cortex

Working memory involves the short-term maintenance of information and is critical in many tasks. The neural circuit dynamics underlying working memory remain poorly understood, with different aspects of prefrontal cortical (PFC) responses explained by different putative mechanisms. By mathematical analysis, numerical simulations, and using recordings from monkey PFC, we investigate a critical but hitherto ignored aspect of working memory dynamics: information loading. We find that, contrary to common assumptions, optimal loading of information into working memory involves inputs that are largely orthogonal, rather than similar, to the late delay activities observed during memory maintenance, naturally leading to the widely observed phenomenon of dynamic coding in PFC. Using a theoretically principled metric, we show that PFC exhibits the hallmarks of optimal information loading. We also find that optimal information loading emerges as a general dynamical strategy in task-optimized recurrent neural networks. Our theory unifies previous, seemingly conflicting theories of memory maintenance based on attractor or purely sequential dynamics and reveals a normative principle underlying dynamic coding.

Exploring weight initialization, diversity of solutions, and degradation in recurrent neural networks trained for temporal and decision-making tasks

Recurrent Neural Networks (RNNs) are frequently used to model aspects of brain function and structure. In this work, we trained small fully-connected RNNs to perform temporal and flow control tasks with time-varying stimuli. Our results show that different RNNs can solve the same task by converging to different underlying dynamics and also how the performance gracefully degrades as either network size is decreased, interval duration is increased, or connectivity damage is induced. For the considered tasks, we explored how robust the network obtained after training can be according to task parameterization. In the process, we developed a framework that can be useful to parameterize other tasks of interest in computational neuroscience. Our results are useful to quantify different aspects of the models, which are normally used as black boxes and need to be understood in order to model the biological response of cerebral cortex areas.

Reconstructing computational system dynamics from neural data with recurrent neural networks

Computational models in neuroscience usually take the form of systems of differential equations. The behaviour of such systems is the subject of dynamical systems theory. Dynamical systems theory provides a powerful mathematical toolbox for analysing neurobiological processes and has been a mainstay of computational neuroscience for decades. Recently, recurrent neural networks (RNNs) have become a popular machine learning tool for studying the non-linear dynamics of neural and behavioural processes by emulating an underlying system of differential equations. RNNs have been routinely trained on similar behavioural tasks to those used for animal subjects to generate hypotheses about the underlying computational mechanisms. By contrast, RNNs can also be trained on the measured physiological and behavioural data, thereby directly inheriting their temporal and geometrical properties. In this way they become a formal surrogate for the experimentally probed system that can be further analysed, perturbed and simulated. This powerful approach is called dynamical system reconstruction. In this Perspective, we focus on recent trends in artificial intelligence and machine learning in this exciting and rapidly expanding field, which may be less well known in neuroscience. We discuss formal prerequisites, different model architectures and training approaches for RNN-based dynamical system reconstructions, ways to evaluate and validate model performance, how to interpret trained models in a neuroscience context, and current challenges.

Training biologically plausible recurrent neural networks on cognitive tasks with long-term dependencies

Training recurrent neural networks (RNNs) has become a go-to approach for generating and evaluating mechanistic neural hypotheses for cognition. The ease and efficiency of training RNNs with backpropagation through time and the availability of robustly supported deep learning libraries has made RNN modeling more approachable and accessible to neuroscience. Yet, a major technical hindrance remains. Cognitive processes such as working memory and decision making involve neural population dynamics over a long period of time within a behavioral trial and across trials. It is difficult to train RNNs to accomplish tasks where neural representations and dynamics have long temporal dependencies without gating mechanisms such as LSTMs or GRUs which currently lack experimental support and prohibit direct comparison between RNNs and biological neural circuits. We tackled this problem based on the idea of specialized skip-connections through time to support the emergence of task-relevant dynamics, and subsequently reinstitute biological plausibility by reverting to the original architecture. We show that this approach enables RNNs to successfully learn cognitive tasks that prove impractical if not impossible to learn using conventional methods. Over numerous tasks considered here, we achieve less training steps and shorter wall-clock times, particularly in tasks that require learning long-term dependencies via temporal integration over long timescales or maintaining a memory of past events in hidden-states. Our methods expand the range of experimental tasks that biologically plausible RNN models can learn, thereby supporting the development of theory for the emergent neural mechanisms of computations involving long-term dependencies.

Cerebellar contributions across behavioural timescales: a review from the perspective of cerebro-cerebellar interactions

Performing successful adaptive behaviour relies on our ability to process a wide range of temporal intervals with certain precision. Studies on the role of the cerebellum in temporal information processing have adopted the dogma that the cerebellum is involved in sub-second processing. However, emerging evidence shows that the cerebellum might be involved in suprasecond temporal processing as well. Here we review the reciprocal loops between cerebellum and cerebral cortex and provide a theoretical account of cerebro-cerebellar interactions with a focus on how cerebellar output can modulate cerebral processing during learning of complex sequences. Finally, we propose that while the ability of the cerebellum to support millisecond timescales might be intrinsic to cerebellar circuitry, the ability to support supra-second timescales might result from cerebellar interactions with other brain regions, such as the prefrontal cortex.

The dynamics and geometry of choice in premotor cortex

The brain represents sensory variables in the coordinated activity of neural populations, in which tuning curves of single neurons define the geometry of the population code. Whether the same coding principle holds for dynamic cognitive variables remains unknown because internal cognitive processes unfold with a unique time course on single trials observed only in the irregular spiking of heterogeneous neural populations. Here we show the existence of such a population code for the dynamics of choice formation in the primate premotor cortex. We developed an approach to simultaneously infer population dynamics and tuning functions of single neurons to the population state. Applied to spike data recorded during decision-making, our model revealed that populations of neurons encoded the same dynamic variable predicting choices, and heterogeneous firing rates resulted from the diverse tuning of single neurons to this decision variable. The inferred dynamics indicated an attractor mechanism for decision computation. Our results reveal a common geometric principle for neural encoding of sensory and dynamic cognitive variables.

Fronto-parietal networks shape human conscious report through attention gain and reorienting

How do attention and consciousness interact in the human brain? Rival theories of consciousness disagree on the role of fronto-parietal attentional networks in conscious perception. We recorded neural activity from 727 intracerebral contacts in 13 epileptic patients, while they detected near-threshold targets preceded by attentional cues. Clustering revealed three neural patterns: first, attention-enhanced conscious report accompanied sustained right-hemisphere fronto-temporal activity in networks connected by the superior longitudinal fasciculus (SLF) II-III, and late accumulation of activity (>300 ms post-target) in bilateral dorso-prefrontal and right-hemisphere orbitofrontal cortex (SLF I-III). Second, attentional reorienting affected conscious report through early, sustained activity in a right-hemisphere network (SLF III). Third, conscious report accompanied left-hemisphere dorsolateral-prefrontal activity. Task modeling with recurrent neural networks revealed multiple clusters matching the identified brain clusters, elucidating the causal relationship between clusters in conscious perception of near-threshold targets. Thus, distinct, hemisphere-asymmetric fronto-parietal networks support attentional gain and reorienting in shaping human conscious experience.

Remapping in a recurrent neural network model of navigation and context inference

Neurons in navigational brain regions provide information about position, orientation, and speed relative to environmental landmarks. These cells also change their firing patterns ('remap') in response to changing contextual factors such as environmental cues, task conditions, and behavioral states, which influence neural activity throughout the brain. How can navigational circuits preserve their local computations while responding to global context changes? To investigate this question, we trained recurrent neural network models to track position in simple environments while at the same time reporting transiently-cued context changes. We show that these combined task constraints (navigation and context inference) produce activity patterns that are qualitatively similar to population-wide remapping in the entorhinal cortex, a navigational brain region. Furthermore, the models identify a solution that generalizes to more complex navigation and inference tasks. We thus provide a simple, general, and experimentally-grounded model of remapping as one neural circuit performing both navigation and context inference.

Natural statistics support a rational account of confidence biases

Previous work has sought to understand decision confidence as a prediction of the probability that a decision will be correct, leading to debate over whether these predictions are optimal, and whether they rely on the same decision variable as decisions themselves. This work has generally relied on idealized, low-dimensional models, necessitating strong assumptions about the representations over which confidence is computed. To address this, we used deep neural networks to develop a model of decision confidence that operates directly over high-dimensional, naturalistic stimuli. The model accounts for a number of puzzling dissociations between decisions and confidence, reveals a rational explanation of these dissociations in terms of optimization for the statistics of sensory inputs, and makes the surprising prediction that, despite these dissociations, decisions and confidence depend on a common decision variable.

Spiking Recurrent Neural Networks Represent Task-Relevant Neural Sequences in Rule-Dependent Computation

Background

Prefrontal cortical neurons play essential roles in performing rule-dependent tasks and working memory-based decision making.

Methods

Motivated by PFG recordings of task-performing mice, we developed an excitatory-inhibitory spiking recurrent neural network (SRNN) to perform a rule-dependent two-alternative forced choice (2AFC) task. We imposed several important biological constraints onto the SRNN, and adapted spike frequency adaptation (SFA) and SuperSpike gradient methods to train the SRNN efficiently.

Results

The trained SRNN produced emergent rule-specific tunings in single-unit representations, showing rule-dependent population dynamics that resembled experimentally observed data. Under varying test conditions, we manipulated the SRNN parameters or configuration in computer simulations, and we investigated the impacts of rule-coding error, delay duration, recurrent weight connectivity and sparsity, and excitation/inhibition (E/I) balance on both task performance and neural representations.

Conclusions

Overall, our modeling study provides a computational framework to understand neuronal representations at a fine timescale during working memory and cognitive control, and provides new experimentally testable hypotheses in future experiments.

Distinctive properties of biological neural networks and recent advances in bottom-up approaches toward a better biologically plausible neural network

Although it may appear infeasible and impractical, building artificial intelligence (AI) using a bottom-up approach based on the understanding of neuroscience is straightforward. The lack of a generalized governing principle for biological neural networks (BNNs) forces us to address this problem by converting piecemeal information on the diverse features of neurons, synapses, and neural circuits into AI. In this review, we described recent attempts to build a biologically plausible neural network by following neuroscientifically similar strategies of neural network optimization or by implanting the outcome of the optimization, such as the properties of single computational units and the characteristics of the network architecture. In addition, we proposed a formalism of the relationship between the set of objectives that neural networks attempt to achieve, and neural network classes categorized by how closely their architectural features resemble those of BNN. This formalism is expected to define the potential roles of top-down and bottom-up approaches for building a biologically plausible neural network and offer a map helping the navigation of the gap between neuroscience and AI engineering.

A unifying perspective on neural manifolds and circuits for cognition

Two different perspectives have informed efforts to explain the link between the brain and behaviour. One approach seeks to identify neural circuit elements that carry out specific functions, emphasizing connectivity between neurons as a substrate for neural computations. Another approach centres on neural manifolds - low-dimensional representations of behavioural signals in neural population activity - and suggests that neural computations are realized by emergent dynamics. Although manifolds reveal an interpretable structure in heterogeneous neuronal activity, finding the corresponding structure in connectivity remains a challenge. We highlight examples in which establishing the correspondence between low-dimensional activity and connectivity has been possible, unifying the neural manifold and circuit perspectives. This relationship is conspicuous in systems in which the geometry of neural responses mirrors their spatial layout in the brain, such as the fly navigational system. Furthermore, we describe evidence that, in systems in which neural responses are heterogeneous, the circuit comprises interactions between activity patterns on the manifold via low-rank connectivity. We suggest that unifying the manifold and circuit approaches is important if we are to be able to causally test theories about the neural computations that underlie behaviour.

Remapping in a recurrent neural network model of navigation and context inference

Neurons in navigational brain regions provide information about position, orientation, and speed relative to environmental landmarks. These cells also change their firing patterns ("remap") in response to changing contextual factors such as environmental cues, task conditions, and behavioral state, which influence neural activity throughout the brain. How can navigational circuits preserve their local computations while responding to global context changes? To investigate this question, we trained recurrent neural network models to track position in simple environments while at the same time reporting transiently-cued context changes. We show that these combined task constraints (navigation and context inference) produce activity patterns that are qualitatively similar to population-wide remapping in the entorhinal cortex, a navigational brain region. Furthermore, the models identify a solution that generalizes to more complex navigation and inference tasks. We thus provide a simple, general, and experimentally-grounded model of remapping as one neural circuit performing both navigation and context inference.

Cortical interneurons: fit for function and fit to function? Evidence from development and evolution

Cortical inhibitory interneurons form a broad spectrum of subtypes. This diversity suggests a division of labor, in which each cell type supports a distinct function. In the present era of optimisation-based algorithms, it is tempting to speculate that these functions were the evolutionary or developmental driving force for the spectrum of interneurons we see in the mature mammalian brain. In this study, we evaluated this hypothesis using the two most common interneuron types, parvalbumin (PV) and somatostatin (SST) expressing cells, as examples. PV and SST interneurons control the activity in the cell bodies and the apical dendrites of excitatory pyramidal cells, respectively, due to a combination of anatomical and synaptic properties. But was this compartment-specific inhibition indeed the function for which PV and SST cells originally evolved? Does the compartmental structure of pyramidal cells shape the diversification of PV and SST interneurons over development? To address these questions, we reviewed and reanalyzed publicly available data on the development and evolution of PV and SST interneurons on one hand, and pyramidal cell morphology on the other. These data speak against the idea that the compartment structure of pyramidal cells drove the diversification into PV and SST interneurons. In particular, pyramidal cells mature late, while interneurons are likely committed to a particular fate (PV vs. SST) during early development. Moreover, comparative anatomy and single cell RNA-sequencing data indicate that PV and SST cells, but not the compartment structure of pyramidal cells, existed in the last common ancestor of mammals and reptiles. Specifically, turtle and songbird SST cells also express the Elfn1 and Cbln4 genes that are thought to play a role in compartment-specific inhibition in mammals. PV and SST cells therefore evolved and developed the properties that allow them to provide compartment-specific inhibition before there was selective pressure for this function. This suggest that interneuron diversity originally resulted from a different evolutionary driving force and was only later co-opted for the compartment-specific inhibition it seems to serve in mammals today. Future experiments could further test this idea using our computational reconstruction of ancestral Elfn1 protein sequences.

Computational modeling of human multisensory spatial representation by a neural architecture

Our brain constantly combines sensory information in unitary percept to build coherent representations of the environment. Even though this process could appear smooth, integrating sensory inputs from various sensory modalities must overcome several computational issues, such as recoding and statistical inferences problems. Following these assumptions, we developed a neural architecture replicating humans' ability to use audiovisual spatial representations. We considered the well-known ventriloquist illusion as a benchmark to evaluate its phenomenological plausibility. Our model closely replicated human perceptual behavior, proving a truthful approximation of the brain's ability to develop audiovisual spatial representations. Considering its ability to model audiovisual performance in a spatial localization task, we release our model in conjunction with the dataset we recorded for its validation. We believe it will be a powerful tool to model and better understand multisensory integration processes in experimental and rehabilitation environments.

Ensemble perspective for understanding temporal credit assignment

Recurrent neural networks are widely used for modeling spatiotemporal sequences in both nature language processing and neural population dynamics. However, understanding the temporal credit assignment is hard. Here, we propose that each individual connection in the recurrent computation is modeled by a spike and slab distribution, rather than a precise weight value. We then derive the mean-field algorithm to train the network at the ensemble level. The method is then applied to classify handwritten digits when pixels are read in sequence, and to the multisensory integration task that is a fundamental cognitive function of animals. Our model reveals important connections that determine the overall performance of the network. The model also shows how spatiotemporal information is processed through the hyperparameters of the distribution, and moreover reveals distinct types of emergent neural selectivity. To provide a mechanistic analysis of the ensemble learning, we first derive an analytic solution of the learning at the infinitely large network limit. We then carry out a low-dimensional projection of both neural and synaptic dynamics, analyze symmetry breaking in the parameter space, and finally demonstrate the role of stochastic plasticity in the recurrent computation. Therefore, our study sheds light on mechanisms of how weight uncertainty impacts the temporal credit assignment in recurrent neural networks from the ensemble perspective.

Adaptive robustness through incoherent signaling mechanisms in a regenerative brain

Animal behavior emerges from collective dynamics of interconnected neurons, making it vulnerable to connectome damage. Paradoxically, many organisms maintain significant behavioral output after large-scale neural injury. Molecular underpinnings of this extreme robustness remain largely unknown. Here, we develop a quantitative behavioral analysis pipeline to measure previously uncharacterized long-lasting latent memory states in planarian flatworms during whole-brain regeneration. By combining >20,000 animal trials with neural population dynamic modeling, we show that long-range volumetric peptidergic signals allow the planarian to rapidly reestablish latent states and restore coarse behavior after large structural perturbations to the nervous system, while small-molecule neuromodulators gradually refine the precision. The different time and length scales of neuropeptide and small-molecule transmission generate incoherent patterns of neural activity which competitively regulate behavior and memory. Controlling behavior through opposing communication mechanisms creates a more robust system than either alone and may serve as a generic approach to construct robust neural networks.

Modelling human behaviour in cognitive tasks with latent dynamical systems

Response time data collected from cognitive tasks are a cornerstone of psychology and neuroscience research, yet existing models of these data either make strong assumptions about the data-generating process or are limited to modelling single trials. We introduce task-DyVA, a deep learning framework in which expressive dynamical systems are trained to reproduce sequences of response times observed in data from individual human subjects. Models fitted to a large task-switching dataset captured subject-specific behavioural differences with high temporal precision, including task-switching costs. Through perturbation experiments and analyses of the models' latent dynamics, we find support for a rational account of switch costs in terms of a stability-flexibility trade-off. Thus, our framework can be used to discover interpretable cognitive theories that explain how the brain dynamically gives rise to behaviour.

Distributed context-dependent choice information in mouse posterior cortex

Choice information appears in multi-area brain networks mixed with sensory, motor, and cognitive variables. In the posterior cortex-traditionally implicated in decision computations-the presence, strength, and area specificity of choice signals are highly variable, limiting a cohesive understanding of their computational significance. Examining the mesoscale activity in the mouse posterior cortex during a visual task, we found that choice signals defined a decision variable in a low-dimensional embedding space with a prominent contribution along the ventral visual stream. Their subspace was near-orthogonal to concurrently represented sensory and motor-related activations, with modulations by task difficulty and by the animals' attention state. A recurrent neural network trained with animals' choices revealed an equivalent decision variable whose context-dependent dynamics agreed with that of the neural data. Our results demonstrated an independent, multi-area decision variable in the posterior cortex, controlled by task features and cognitive demands, possibly linked to contextual inference computations in dynamic animal-environment interactions.